Automatic Analysis Pipeline of Squatting Ability for Elderly People Based on An Intelligent Platform

As people age, various functions of the human body will decline gradually, leading to an increased risk of motor dysfunction. The assessment of motor function in the elderly has a great significance for maintaining their health and improving their ability to live independently. In this work, we propose an automatic analysis pipeline of squatting ability for elderly people based on computer vision and machine learning methods. With our intelligent application platform designed for assessing and monitoring the health status of elder individuals, they could capture motor function videos and achieve the corresponding analysis and assessment results. To validate the accuracy of our motor function parameter calculation algorithms, we conducted a multi-modal experiment including the 3D optical motion capture and video recording with mobile phones. Our algorithms reached a high correlation of 0.86 with results from 3D optical motion capture system. We also conducted an outdoor experiment in the community to explore the statistical performance of motor functions of elderly people. The pipeline proposed in this work can help elderly people receive remote motor function analysis, and provide them with a low-cost, high-efficiency health monitoring and long-term tracking tool.